RE-GrievanceAssist: Enhancing Customer Experience through ML-Powered Complaint Management
Venkatesh C, Harshit Oberoi, Anurag Kumar Pandey, Anil Goyal, Nikhil Sikka
TL;DR
RE-GrievanceAssist tackles the challenge of high-volume real estate customer complaints by delivering an end-to-end ML pipeline for triage, routing, and auto-replies. It combines a response/no-response classifier using TF-IDF with XGBoost, a user-type classifier using FastText, and a hierarchical issue/sub-issue classifier using TF-IDF plus XGBoost, deployed as a Databricks batch workflow. On a test set, the components achieve F1-scores of $86.39\%$, $90\%$, $72.95\%$, and $62.23\%$ for response/no-response, user-type, issue, and sub-issue respectively, indicating strong discriminative performance. In deployment, the system automates around $40\%$ of tickets and reduces manual effort by about $50\%$ for the remainder, yielding monthly savings of Rs $1,50,000$ since August 2023 and facilitating faster grievance resolution in the real-estate domain.
Abstract
In recent years, digital platform companies have faced increasing challenges in managing customer complaints, driven by widespread consumer adoption. This paper introduces an end-to-end pipeline, named RE-GrievanceAssist, designed specifically for real estate customer complaint management. The pipeline consists of three key components: i) response/no-response ML model using TF-IDF vectorization and XGBoost classifier ; ii) user type classifier using fasttext classifier; iii) issue/sub-issue classifier using TF-IDF vectorization and XGBoost classifier. Finally, it has been deployed as a batch job in Databricks, resulting in a remarkable 40% reduction in overall manual effort with monthly cost reduction of Rs 1,50,000 since August 2023.